Designing Pressure Vessels for Nozzle Loads

Many years ago, when the calculation report of a pressure vessels (doesn’t matter according to which pressure vessel standard) was composed by a couple of pages, nobody was really worried about nozzle loads: or better, we all knew that piping connected to the vessel could transmit to the vessel wall some kind of loads: the only problem is that the methods to calculate the stresses due to these loads still did not exist. This is the reason why the ASME or DIN standards for nozzle flanges were made in such a way, that the bolting area provided was 2 or 3 times the area needed to guarantee the tightness of the joint: just for the purpose of giving to the connection an additional resistance to take into account the presence of such (unknown) loads.

Maybe it was just a coincidence, but the fact is that all the engineering companies started to consider nozzle loads in their specifications when the first Welding Research bulletins dealing with calculation of stresses induced by nozzle loads appeared. Personally, I do not remember reports concerning catastrophic accidents caused by nozzle loads, as it was for the case of wind and seismic loads, at that time also not yet provided by the pressure vessel standards (but at least considered by the standards for building). I only remember that the first approach followed by the great majority of the engineering companies was first of all to place orders on all the pressure vessels and heat exchangers giving only the design temperature and the design pressure: than they could proceed with the calculation of piping, and then they were able to give the manufacturer the nozzle loads to be considered. Unfortunately, consideration of these loads based on the new calculation methods caused in many cases an increase in the wall thicknesses already calculated by the manufacturers, with the need to modify orders of materials already placed, in many cases when the materials had been already delivered to the shop. This of course was a good reason for the manufacturers to claim important extra costs: everybody can easily understand that the evaluation of such unavoidable modifications is sometimes made with a more severe criterion than the one used in preparing an offer, where you know that you are in competition with other companies, and therefore you must keep your price as low as possible in order to be able to get the order. But once you have got the order, for order modifications and extra costs there is no competitor…

Due to these reasons, the engineering companies decided to change their politics on the subject of nozzle loads: therefore, since it was not possible to run the calculation of piping without having already the fabrication drawings of the vessels, they prepared tables with specified standard nozzle loads (forces and moments), different for the different nozzle diameters, and generally with still undefined directions, asking the manufacturers to calculate the wall thicknesses of the vessels taking into account these specified figures. The result of this change of politics was a general increase of the thicknesses and therefore of the prices of the vessels: also because the new methods were very good to give a reasonably correct mathematical calculation of the stresses caused by these presumed loads, but failed to give answers for some particular problems, such as the composition of such stresses with the local stresses caused by pressure existing close to the opening, and the determination of the categories of the stresses from the point of view of the design by analysis: primary or secondary? membrane or bending? local or general? And at the end, what is the nature of these specified nozzle loads? Are they thermal stresses caused by restrained thermal expansion of piping, or mechanical stresses due to weight, wind, seism? Moreover, how are these loads calculated? Is the elasticity of the wall considered, or the nozzle is regarded as a fixed point without the possibility of any deformation? Clearly, it is not possible to give the proper answer to these questions with the use of a standard table of nozzle loads, which obliges the designer to consider the worst possible answer in all cases.

And at the end, are we sure that all this is not a purely theoretical exercise? Well, in the more than 50 years of my professional life I have heard a lot of discussions on nozzle load calculations, but, as I said before, I have never seen neither severe accidents, nor even sensible damages caused by these phantomatic loads: moreover, as European Standards started to invent new methods for nozzle load calculations, the difference among such methods became more and more evident: the maximum axial force, or the maximum bending moment supported by a nozzle could become two or three times higher, or lower, if calculated with different standards: with important variations due to the amount of pressure existing in the vessel. Sometimes the designers of the engineering companies, responsible for checking manufacturers’ calculations, seem to be particularly worried not only about the loads, but also about the reactions induced in the vessel supports by such loads. Of course this is certainly right: any load applied to a nozzle will generate stresses not only at the nozzle location, but also at the vessel supports. But when the loads you are considering are not the result of a complete piping analysis, because they are just conventional figures that the manufacturer is obliged to take into account in the determination of the vessel thickness (with the additional complication that the specification doesn’t indicate the real direction of the loads, but simply tells the manufacturer to consider the worst possible direction), can you tell me how is it possible to calculate the reactions at the supports of the vessel? Just to make a stupid example, what should happen to a distillation column containing 50 or 60 nozzles in the case that the conventional nozzle loads are all oriented towards the same direction? Well, of course this is a practical aspect of a more general problem, particularly important in the case of tall vertical vessels: that is, to convert a loading capacity into an actual load: for example, when you design a platform in a column, you generally consider a design load of (maybe) 2500 N/m², but this doesn’t mean that all the platforms of that column must be fully loaded with that load, which would have serious consequences also at the foundations. Therefore, most pressure vessel standards (and also most building standards) try to give an answer to the problem, by specifying the load combinations to be taken into account: however, in the case of nozzle loads, very seldom you will find detailed instructions in the pressure vessel standards. One of the standards that seems to have considered the problem is the European Harmonized standard EN 13445.3:2021. In Clause 22 (Static Analysis of Tall Vertical Vessels on Skirts) you find the following sentences:

“Additional forces from attached piping, other than weight, wind and earthquake loads, shall be considered, see …….. It is the responsibility of the designer to decide to what extent additional forces from attached piping shall be taken into account for the static analysis of columns since their influence depends on the whole behaviour of the column and piping configuration (see NOTE ………).”

And then:

”In the case that multiple pipes are connected to the column the resulting horizontal reaction forces and their directions shall be vector combined at each elevation taking into account the direction of each of the single pipe forces. Where actual forces and their directions are not available it is not reasonable to assume that all horizontal forces act in the same direction. The maximum resulting shear force at the base of the column shall be vector combined from the horizontal resulting forces and their directions of all elevations. The maximum resulting bending moment at the base of the column shall be vector combined from the moments and their directions determined from these horizontal resulting forces with their directions and elevations”.

In other words: the decision belongs to the designer, who should be a reasonable person with a lot of commonsense. Well, the problem is that in Pressure Vessel Design there is always a designer (working for the manufacturer) and a controller (working for the engineering company who has purchased the vessel). The problem is that not always these two guys have the same degree of commonsense.

At the end, the stress report of any pressure vessel (tall or not tall) must contain the loads at the supports: these loads are of course important not only for the calculation of the stresses in the vessel itself, but also for the design of the foundations. Therefore, it is important that anything not clearly defined in the standard should clearly defined at least in the design specifications, in order to avoid unnecessary and time-consuming discussions.

Any comment to this paper will be very much appreciated.

   Fernando Lidonnici           

  Convenor of WG53/CEN TC54

      Milano, July 25^th, 2025